# AI workflow decision record

By Sajad & Hassan, Ampliflow
https://ampliflow.ai/ai-consulting

A blank worksheet for one business workflow. This is a planning aid, not a project proposal, a promised service deliverable or a prediction of results. Complete it locally. Keep customer records, confidential information and access credentials out of this document.

## 1. Name the outcome

- What should happen when the work is done?
- Who needs the result, and who owns the decision?
- What currently prevents that outcome?

## 2. Follow the work

- What starts the process?
- What information is needed? Where does it come from?
- Who handles each step, and who takes over next?
- Where does the work wait, get repeated or lose context?
- What happens when someone is away or information is missing?

## 3. Record a starting point

- Measurement period and source:
- Amount of work received and completed:
- Time spent, outstanding work and errors:
- What is not measured or remains uncertain:

Keep the units and time periods comparable. The smallest percentage in a funnel is not automatically its constraint.

## 4. Compare realistic options

For each option, record the evidence, constraints, running costs and the smallest test that could change your mind.

- Change the process or clarify ownership.
- Configure a feature in software you already use.
- Use a fixed rule for a predictable step.
- Use AI assistance for varied information, with appropriate human checks.
- Investigate a new build, or leave the process unchanged.

## 5. Define a bounded test

- Normal examples to check:
- Missing, ambiguous or conflicting inputs:
- Permissions and actions that must remain under human control:
- What should happen on a failed step, duplicate request or unavailable system:
- How repeated outputs will be checked:
- What evidence would make you continue, change direction or stop:

## 6. Decide who carries it forward

- Decision owner and people affected:
- Access, implementation, training and support still to agree:
- Who reviews the result, and when:
- Result compared with the starting point:
- What changed, what did not, and what remains uncertain:

Measure the business outcome separately from the quality of an AI answer. Record observations before claiming an improvement.

## Where to go next

- AI consulting: https://ampliflow.ai/ai-consulting
- Research and feasibility: https://ampliflow.ai/services/research-feasibility
- Business automation: https://ampliflow.ai/services/automation
- Workplace AI learning: https://ampliflow.ai/services/ai-training/workplace
- Contact: hello@ampliflow.ai